AAN update 2017 DEC.pdf

SPECIAL ARTICLE
LEVEL OF RECOMMENDATION
Practice guideline update summary: Mild
cognitive impairment
Report of the Guideline Development, Dissemination, and Implementation
Subcommittee of the American Academy of Neurology
Ronald C. Petersen, MD, PhD, Oscar Lopez, MD, Melissa J. Armstrong, MD, MSc, Thomas S.D. Getchius,
Correspondence
Mary Ganguli, MD, MPH, David Gloss, MD, MPH&TM, Gary S. Gronseth, MD, Daniel Marson, JD, PhD,
American Academy of
Tamara Pringsheim, MD, Gregory S. Day, MD, MSc, Mark Sager, MD, James Stevens, MD, and
Neurology
Alexander Rae-Grant, MD
[email protected]
Neurology® 2018;90:1-10. doi:10.1212/WNL.0000000000004826
Abstract
MORE ONLINE
Objective
Podcast
To update the 2001 American Academy of Neurology (AAN) guideline on mild cognitive
Dr. Jeﬀ Burns talks with
impairment (MCI).
Dr. Ronald Petersen about
the updated AAN guideline
Methods
on mild cognitive
The guideline panel systematically reviewed MCI prevalence, prognosis, and treatment articles
impairment.
according to AAN evidence classiﬁcation criteria, and based recommendations on evidence and
Npub.org/ojn0w9
modiﬁed Delphi consensus.
Results
MCI prevalence was 6.7% for ages 60–64, 8.4% for 65–69, 10.1% for 70–74, 14.8% for 75–79,
and 25.2% for 80–84. Cumulative dementia incidence was 14.9% in individuals with MCI older
than age 65 years followed for 2 years. No high-quality evidence exists to support pharmaco-
logic treatments for MCI. In patients with MCI, exercise training (6 months) is likely to
improve cognitive measures and cognitive training may improve cognitive measures.
Major recommendations
Clinicians should assess for MCI with validated tools in appropriate scenarios (Level B).
Clinicians should evaluate patients with MCI for modiﬁable risk factors, assess for functional
impairment, and assess for and treat behavioral/neuropsychiatric symptoms (Level B). Clini-
cians should monitor cognitive status of patients with MCI over time (Level B). Cognitively
impairing medications should be discontinued where possible and behavioral symptoms treated
(Level B). Clinicians may choose not to oﬀer cholinesterase inhibitors (Level B); if oﬀering,
they must ﬁrst discuss lack of evidence (Level A). Clinicians should recommend regular
exercise (Level B). Clinicians may recommend cognitive training (Level C). Clinicians should
discuss diagnosis, prognosis, long-term planning, and the lack of eﬀective medicine options
(Level B), and may discuss biomarker research with patients with MCI and families (Level C).
From the Department of Neurology (R.C.P.), Mayo Clinic, Rochester, MN; Department of Neurology (O.L.), University of Pittsburgh Medical Center, PA; Department of Neurology
(M.J.A.), University of Florida College of Medicine, Gainesville; Heart Rhythm Society (T.S.D.G.), Washington, DC; Department of Psychiatry (M.G.), University of Pittsburgh, PA;
Department of Neurology (D.G.), Charleston Area Medical Center, WV; Department of Neurology (G.S.G.), University of Kansas Medical Center, Kansas City; Department of Neurology
(D.M.), University of Alabama, Birmingham; Department of Clinical Neurosciences, Psychiatry, Pediatrics and Community Health Sciences (T.P.), Cumming School of Medicine,
University of Calgary, Canada; Knight Alzheimer Disease Research Center (G.S.D.), Washington University School of Medicine, St. Louis, MO; Wisconsin Alzheimer’s Institute (M.S.),
School of Medicine and Public Health, University of Wisconsin, Madison; Department of Neurology (J.S.), Fort Wayne Neurological Center, IN; and Department of Neurology (A.R.-G.),
Cleveland Clinic, OH.
Approved by the Guideline Development, Dissemination, and Implementation Subcommittee on July 16, 2016; by the Practice Committee on August 22, 2016; and by the AAN Institute
Board of Directors on October 5, 2017.
This guideline was endorsed by the Alzheimer’s Association on May 1, 2017.
Go to Neurology.org/N for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.
Copyright © 2017 American Academy of Neurology
1
Copyright ª 2017 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Glossary
AAN = American Academy of Neurology; AD = Alzheimer disease; aMCI = amnestic mild cognitive impairment; CI =
conﬁdence interval; CIND = cognitively impaired, no dementia; FDA = Food and Drug Administration; IADL = instrumental
activities of daily living; MCI = mild cognitive impairment; RR = relative risk.
Mild cognitive impairment (MCI) is a condition in which
Description of the analytic process
individuals demonstrate cognitive impairment with minimal
impairment of instrumental activities of daily living (IADL).1–3
This practice guideline principally follows the methodolo-
Although MCI can be the ﬁrst cognitive expression of Alz-
gies described in the 2004 edition of the AAN’s guideline
heimer disease (AD), it can also be secondary to other disease
development process manual.7 Conclusions and recom-
processes (i.e., other neurologic, neurodegenerative, systemic,
mendations were developed in accordance with the process
or psychiatric disorders).4 The term amnestic MCI (aMCI)
outlined in the 2011 guideline development process man-
describes a syndrome in which memory dysfunction predom-
ual, as amended to include the updated scheme for classi-
inates; in nonamnestic MCI, impairment of other cognitive
fying therapeutic articles.8 The complete guideline provides
features (e.g., language, visuospatial, executive) is more
a description of the exact methodology followed, including
prominent.2
the processes of convening the author panel, performing the
literature search, and reviewing the evidence. In accordance
This practice guideline updates a 2001 American Academy of
with the 2011 guideline manual, recommendations were
Neurology (AAN) practice parameter with recommendations
based not only on the evidence in the systematic review, but
concerning the diagnosis and treatment of MCI.5 The guide-
also on strong related evidence, established principles of
line focuses on presumed idiopathic or neurodegenerative
care, and inferences. The level of obligation for each rec-
MCI—particularly relating to AD—rather than mild cognitive
ommendation was based on the strength of these premises
changes relating to potentially reversible causes (e.g., metabolic,
and the risk–beneﬁt ratio of following the recommendation,
vascular, systemic, or psychiatric disorders) or Parkinson
with adjustments based on importance of outcomes, vari-
disease–MCI or vascular cognitive impairment, as these may have
ation in patient preferences, feasibility/availability, and
diﬀerent epidemiologic and treatment spectra than AD. This
patient costs. Consensus was determined by a modiﬁed
article summarizes the guideline ﬁndings, conclusions, and rec-
Delphi voting process in accordance with prespeciﬁed
ommendations. The full text of the guideline, including appen-
rules.8
dices e-1 through e-8, is available as supplemental data (links.lww.
com/WNL/A125), as are tables e-1 through e-3 (links.lww.com/
The panelists noted that various deﬁnitions of MCI and of
WNL/A34) and references e1–e50 (links.lww.com/WNL/A50).
related terms, such as cognitively impaired, no dementia
(CIND), were used in the reviewed literature. Variation was
The guideline addresses 4 questions:
based on diﬀerent ascertainment methods, diﬀerent neu-
ropsychological measures, diﬀerent measure thresholds,
1.
What is the prevalence of MCI in the general population?
and requirements for diﬀerent cognitive deﬁcits. There was
2.
What is the prognosis for patients diagnosed with MCI
also variation in the use of aMCI and nonamnestic MCI in
for progression to a diagnosis of dementia, and how does
these studies. To address these discrepancies, the panelists
this compare with an age-matched general population?
reﬂected the speciﬁc deﬁnition used for a study where fea-
3.
What pharmacologic treatments are eﬀective for patients
sible in the evidence synthesis tables and guideline text, and
diagnosed with MCI?
provided speciﬁc comments on the potential eﬀect of dif-
4.
What nonpharmacologic treatments are eﬀective for
fering deﬁnitions.
patients diagnosed with MCI?
Analysis of evidence
This guideline does not review the rapidly evolving ﬁeld of bio-
marker research in MCI; the guideline panel determined that this
What is the prevalence of MCI in the
should be the subject of a future guideline or systematic review. In
general population?
addition, the potential psychological distress of a diagnosis of
Background
MCI (which has been discussed in the literature) was not one of
Various deﬁnitions of MCI have been used over time,
the questions reviewed by the expert panel for this guideline.6
reﬂecting an evolution of thought from primarily focusing on
amnesia to including other cognitive deﬁcits. Because
Supplemental Data
memory deﬁcits are the clinical hallmark of AD, some groups
Full text of guideline at:
used criteria for MCI that required the presence of memory
NPub.org/4evlhy
deﬁcits in isolation (e.g., aMCI),3,9,10 and others included
a broader deﬁnition that included either single-domain
2
Neurology | Volume 90, Number 3 | January 16, 2018
Neurology.org/N
Copyright ª 2017 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

nonamnestic deﬁcits or deﬁcits in multiple cognitive
participants with MCI. However, 2 studies documented
domains, either with memory impairment (multidomain
increased rates of ultimate conversion to dementia in
aMCI) or without (multidomain nonamnestic MCI).1,2,11
participants with MCI who reverted to normal cognition,
The deﬁnition of MCI is also aﬀected by the psychometric
suggesting that individuals who revert remain at a higher
properties of, and norms for, the tests used to identify
risk of progression back to MCI or dementia than indi-
thresholds between normal aging and MCI. Table e-1 (links.
viduals who have never received an MCI diagnosis (in
lww.com/WNL/A34) presents the characteristics of various
these studies, 65%e5 and 55% ultimately converted to
deﬁnitions of MCI used in the literature evaluated here. Table
dementiae8).
e-2 shows the eﬀect on frequency of MCI in the population
when less or more stringent MCI criteria were applied.
Conclusions
Persons with MCI are at higher risk of progressing to dementia
Analysis
than age-matched controls (high conﬁdence, multiple con-
Twenty Class I studies9,10,12–29 and 14 Class II studies30–40,e1–e3
cordant Class I studies, meta-analysis). Persons diagnosed with
were identiﬁed. Eight of the Class I studies showed that
MCI may remain stable, return to neurologically intact, or
a lower education level was signiﬁcantly associated with
progress to dementia (multiple Class I studies, 14.4%–55.6%
a higher prevalence of MCI.9,10,14,18,21,24,27,28 Two of the Class I
reverting to normal).
studies indicated that male sex was associated with the pres-
ence of MCI13,24 but other studies found similar baseline
prevalence in men and women.14,15,27
What pharmacologic treatments are available
for patients diagnosed with MCI, and are these
A random-eﬀects meta-analysis using Class I and II studies
treatments effective on cognitive measures of
conﬁrmed an increased prevalence with cohort age. The all-
progression to dementia, excluding other
studies estimate for individuals aged 60–64 years was 6.7%
symptomatic effects?
(95% conﬁdence interval [CI] 3.4%–12.7%, I2 11.0); for
Analysis
those aged 65–69, 8.4% (95% CI 5.2%–13.4%, I2 0); for
One Class I study,e9 10 Class II studies described in 9 pub-
ages 70–74, 10.1% (95% CI 7.5%–13.5%, I2 5.2); for ages
lications,e10–e18 and 3 Class III studiese19–e21 addressed phar-
75–79, 14.8% (95% CI 10.1%–21.1%, I2 60.7); and for ages
macologic treatment of MCI. Table 1 describes the available
80–84, 25.2% (95% CI 16.5%–36.5%, I2 0) (see table e-3,
studies and conclusions for each pharmacologic intervention.
links.lww.com/WNL/A34).
Comprehensive descriptions of each study, including ef-
fect sizes and CIs, are available in the full-length guideline
Conclusions
(links.lww.com/WNL/A125).
MCI is common in older populations, and its prevalence
increases with age (high conﬁdence, multiple Class I and
What nonpharmacologic treatments are
Class II studies, consistent meta-analysis) and lower edu-
effective for patients diagnosed with MCI?
cational level (high conﬁdence, multiple Class I studies).
Analysis
What is the prognosis for patients diagnosed
Two Class II studies were reviewed that used exercise as an
with MCI for progression to a diagnosis of
intervention in individuals with MCI,e22,e23 and 1 Class IIe24
dementia, and how does this compare with an
and 4 Class III studiese25–e28 investigated the use of various
age-matched general population?
cognitive interventions. Table 2 describes the available
studies and conclusions for each nonpharmacologic in-
Analysis
tervention; details are provided in the full-length guideline
Nine Class I studies evaluated prognosis for individuals
(links.lww.com/WNL/A125).
with MCI,9,13,19,23,27,e2,e4–e7 all showing an increased risk of
progression to dementia when participants with MCI were
compared with age-matched participants without MCI. A
Putting the evidence into
random-eﬀects meta-analysis demonstrated that the cumulative
clinical context
incidence for the development of dementia in individuals with
MCI/CIND older than age 65 followed for 2 years was 14.9%
Care for persons with cognitive impairment meeting various
(95% CI 11.6%–19.1%, I2 = 0). In those with MCI/CIND vs
MCI criteria continues to evolve, with the area of biomarker
age-matched participants at 2–5 years after, the relative risk
research changing particularly rapidly. Even in the context of
(RR) of dementia (all types) was 3.3 (95% CI 2.5–4.5, I2 = 4.9);
an evolving ﬁeld, clinicians can provide high-quality care
the RR of the diagnosis of AD was 3.0 (95% CI 2.1–4.8,
focusing on counseling, treatment, and comorbidity man-
I2 = 17.3).
agement. Where clinicians are not proﬁcient in caring for
the cognitive or behavioral/psychiatric needs of persons
Reversion to normal cognition in individuals with MCI
with MCI, referral to appropriate specialists is an important
Four Class I studies9,19,23,e5 showed reversion to normal cog-
part of the treatment paradigm in line with the following
nition on follow-up in 14.4%,19 33.3%,9 19%,23 and 38%e5 of
recommendations.
Neurology.org/N
Neurology | Volume 90, Number 3 | January 16, 2018
3
Copyright ª 2017 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Table 1 Evidence and conclusions for pharmacologic treatments for mild cognitive impairment (MCI)
Agent
Classification of evidence
Conclusion
Donepezil
3 Class II studies (Petersen 2005,e10
In patients with MCI, donepezil use over 3 years is possibly ineffective for reducing the
Doody 2009,e11 Salloway 2004e12)
chances of a progression to possible or probable Alzheimer dementia (low confidence in
the evidence, single Class II study [Petersen 2005e10]). In patients with MCI, it is unknown
whether donepezil slows progression on various cognitive scales (very low confidence in
the evidence based on 2 Class II studies with limited precision and small magnitude of
effect) (Doody 2009,e11 Salloway 2004e12). Study CIs could not exclude an important
effect and the ADAS-Cog change was statistically significant but not clinically meaningful.
Galantamine
2 Class II studies (Winblad 2008,e13 both
In patients with MCI, galantamine use over 24 months is probably ineffective for
studies reported in 1 article)
reducing progression to dementia (moderate confidence in the evidence based on 2
Class II studies).
Rivastigmine
1 Class II study (Feldman 2007e14)
In patients with MCI, rivastigmine use up to 48 months is possibly ineffective for reducing
the rate of progression to possible or probable Alzheimer dementia (low confidence in
the evidence based on a single Class II study).
Flavonoid-
1 Class II study (Desideri 2012e15)
In patients with MCI, there is insufficient evidence to support or refute the cognitive
containing drink
benefits of a drink with high-dose flavonoids (about 990 mg) on an integrated measure
(cognitive z score) of overall cognitive function at 8 weeks (very low confidence in the
evidence based on a single Class II study with CIs including unimportant effects;
evidence of a dose response was also unclear).
Homocysteine-
1 Class II study (Smith 2010e16)
In patients with MCI, there is insufficient evidence to support or refute the use of
lowering B vitamins
homocysteine-lowering therapies in patients with MCI (very low confidence in the
evidence based on a single Class II study with decreased confidence in the evidence
owing to use of a primary endpoint with unclear clinical significance).
Transdermal
1 Class I study (Newhouse 2012e9)
Six months of transdermal nicotine (15 mg/d) use possibly improves cognitive test
nicotine patch
performance but not Clinical Global Impression of Change in patients with aMCI who do
not smoke (low confidence in the evidence based on 1 Class I study with decreased
confidence in the evidence owing to uncertain clinical significance of the outcome of hit
reaction time).
Piribedil
1 Class III study (Nagaraja 2001e19)
Data are insufficient to support or refute an effect of piribedil on cognitive measures in
MCI (very low confidence in the evidence based on 1 Class III study).
Rofecoxiba
1 Class II study (Thal 2005e17)
Rofecoxib possibly increases the risk of progression to AD in patients with MCI (low
confidence in the evidence based on 1 Class II study).
Tesamorelin
1 Class II study (Baker 2012e18)
In patients with MCI, treatment with tesamorelin injections over 20 weeks is possibly
injections
effective to improve performance on various cognitive measures (low confidence in the
evidence based on 1 Class II study).b
V0191
1 Class III study (Dubois 2012e20)
Data are insufficient to support or refute an effect of V0191 use on ADAS-Cog response
rates in patients with MCI (very low confidence in the evidence based on 1 Class III study).
Vitamin E
1 Class II study (Petersen 2005e10)
In patients with MCI, use of vitamin E 2,000 IU daily is possibly ineffective for reducing
progression to AD (low confidence in the evidence based on a single Class II study).
Vitamin E +
1 Class III study (Naeini 2014e21)
In patients with MCI, combined use of oral vitamin E 300 mg and C 400 mg daily over 12
vitamin C
months is of uncertain efficacy (very low confidence in the evidence based on 1 Class III
study).
Abbreviations: AD = Alzheimer disease; ADAS-Cog = Alzheimer’s Disease Assessment Scale–cognitive subscale; aMCI = amnestic mild cognitive impairment;
CI = confidence interval.
References cited here can be found in the e-references (links.lww.com/WNL/A50) for the guideline summary article.
a Rofecoxib was removed from the market worldwide in September 2004. There are no data on whether other anti-inflammatory medications are effective or
harmful in patients with MCI.
b It is unclear from this study whether this is effect is sustained beyond 20 weeks.
Practice recommendations
suggesting that this condition reﬂects a pathologic disease state
rather than normal cognitive aging. Appropriate diagnosis of
Section A: Recommendations for assessing
MCI is important in order to assess for reversible causes of
for MCI
cognitive impairment, to help patients and families understand
Recommendation A1
the cause of their cognitive concerns, and to discuss the
prognostic possibilities with the provider so they can plan ac-
Rationale
cordingly, although sharing the diagnosis must be balanced
Appropriate diagnosis of MCI is important because
with the potential harm of anxieties from diagnosing a patient
MCI becomes increasingly common as individuals age and is
with a condition that may not progress. Ascribing cognitive
associated with an increased risk of progression to dementia,
symptoms to normal aging without an assessment for MCI may
4
Neurology | Volume 90, Number 3 | January 16, 2018
Neurology.org/N
Copyright ª 2017 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Table 2 Evidence and conclusions for nonpharmacologic treatments for mild cognitive impairment (MCI)
Agent
Classification of evidence
Conclusion
Exercise
2 Class II studies (Nagamatsu 2012,e22 Suzuki 2013e23)
In patients with MCI, treatment with exercise training for 6 months is likely to
improve cognitive measures (moderate confidence in the evidence based on
2 Class II studies).
Cognitive
1 Class II (Kinsella 2009e24) and 4 Class III studies (Kinsella
There is insufficient evidence to support or refute the use of any individual
interventions
2016,e25 Tsolaki 2011,e26 Nakatsuka 2015,e27 Lam
cognitive intervention strategy (1 Class II study with results that are not
2015e28)
statistically significant and with suspected imprecision, 4 Class III studies,
each examining a different cognitive intervention strategy). When various
cognitive interventions are considered as a group, for patients with MCI,
cognitive interventions may improve select measures of cognitive function
(low confidence in the evidence based on 1 Class II study with insufficient
precision [Kinsella 2009e24], 1 Class III study showing improvements in
strategy knowledge, internal strategy use, and well-being but not external
strategy or memory [Kinsella 2016e25], 1 Class III study [Tsolaki 2011e26]
showing improvement on multiple cognitive measures, 1 Class III study
[Nakatsuka 2015e27] showing improvement on the MMSE but with some
limitations, and 1 Class III study [Lam 2015e28] showing no differences when
all patients with MCI are considered, but with improvements in the
integrated cognitive–physical training groups when considering the ADAS-
Cog, fluency, and recall in patients with single-domain MCI and fluency in
patients with multidomain MCI).
Abbreviations: ADAS-Cog = Alzheimer’s Disease Assessment Scale–cognitive subscale; MMSE = Mini-Mental State Examination.
References cited here can be found in the e-references (links.lww.com/WNL/A50) for the guideline summary article.
result in failure to assess for reversible causes of cognitive im-
Recommendation A3
pairment or to provide patients and families with an accurate
diagnosis that may aﬀect life choices, or both. Although sub-
Rationale
jective cognitive complaints alone are insuﬃcient to diagnose
When screening or assessing for MCI, validated assessment
MCI,e29 such complaints from either patients or their close
tools should be used. Various instruments have acceptable
contacts are core to most major MCI diagnostic criteria, as they
diagnostic accuracy for detecting MCI, with no instrument
may reﬂect a change in cognitive function.e30
being superior to another.e32 Because brief cognitive as-
sessment instruments are usually calibrated to maximize
sensitivity rather than speciﬁcity, patients who test positive
Recommendation
for MCI should then have further assessment (e.g., more in-
For patients for whom the patient or a close contact voices
depth cognitive testing, such as neuropsychological testing
concern about memory or impaired cognition, clinicians
with interpretation based on appropriate normative data) to
should assess for MCI and not assume the concerns are re-
formally assess for this diagnosis. Diagnosis of MCI is based
lated to normal aging (Level B).
ultimately on a clinical evaluation determining cognitive
function and functional status and not solely on a speciﬁc
Recommendation A2
test score.
Rationale
Recommendation
In the United States, the Medicare Annual Wellness Visit
For patients for whom screening or assessing for MCI is ap-
requires an assessment to detect cognitive impairment.e31
propriate, clinicians should use validated assessment tools to
Subjective cognitive complaints alone can result in over-
assess for cognitive impairment (Level B). For patients who
diagnosis or underdiagnosis of MCI and thus are insuﬃcient
test positive for MCI, clinicians should perform a more formal
to screen for MCI.e29 Clinicians assessing for cognitive im-
clinical assessment for diagnosis of MCI (Level B).
pairment should use a brief, validated cognitive assessment
instrument in addition to eliciting patient and informant
Recommendation A4
history regarding cognitive concerns.
Rationale
In the presence of cognitive impairment, clinicians need to
Recommendation
distinguish between a diagnosis of MCI and one of dementia,
When performing a Medicare Annual Wellness Visit, clini-
although the boundary is not always clear. Diagnosing
cians should not rely on historical report of subjective
dementia prematurely can lead to negative consequences
memory concerns alone when assessing for cognitive im-
for patients and families. Only a proportion of patients with
pairment (Level B).
MCI will proceed to dementia. In patients with cognitive
Neurology.org/N
Neurology | Volume 90, Number 3 | January 16, 2018
5
Copyright ª 2017 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

impairment, clinicians must carefully assess for evidence of
use of biomarkers in patients with MCI is a rapidly evolving
functional impairment limiting independence in daily activi-
ﬁeld,e34–e36 but to date, there are no biomarkers clearly shown
ties (e.g., by taking a careful history from the patient and
to predict progression in patients with MCI.e37
a close contact), a requirement for all dementia diagnoses, to
Recommendation A7a
help distinguish between MCI and dementia. With a speciﬁc
inquiry about functional impairment, clinicians may also
For patients and families asking about biomarkers in MCI,
identify dementia in patients when patients and family are less
clinicians should counsel that there are no accepted bio-
forthcoming about functional problems.
markers available at this time (Level B).
Recommendation
Recommendation A7b
For patients with MCI, clinicians should assess for the pres-
For interested patients, clinicians may discuss the option of
ence of functional impairment related to cognition before
biomarker research or refer patients, or both, if feasible, to centers
giving a diagnosis of dementia (Level B).
or organizations that can connect patients to this research (e.g.,
subspecialty centers, Trial Match, ClinicalTrials.gov) (Level C).
Recommendation A5
Recommendation A8
Rationale
Rationale
Diagnoses of MCI and dementia have important implications
for patients and families. Appropriate diagnosis is important for
Because patients with MCI can improve, remain stable, or
informing evaluation for underlying causes, counseling on long-
progress cognitively over time, patients must be monitored
term prognosis, and recommending therapeutic strategies.
serially for changes in status that could change diagnosis and
Clinicians in many disciplines can have experience in caring for
thus management approach (e.g., treatment, counseling).
individuals with cognitive impairment, including family prac-
Although MCI has no approved pharmacologic management,
tice, geriatrics, internal medicine, neurology, psychiatry, and
there are US Food and Drug Administration (FDA)–
psychology. When clinicians without experience in cognitive
approved agents for treatment of Alzheimer dementia,e38–e42
impairment identify patients for whom there is a concern of
further emphasizing the importance of assessing for a change
MCI, they should refer these patients to a specialist with ex-
in cognitive status over time.
perience in cognition for further evaluation.
Recommendation
Recommendation
For patients suspected to have MCI, clinicians who lack the
For patients diagnosed with MCI, clinicians should perform
necessary experience should refer these patients to a specialist
serial assessments over time to monitor for changes in cog-
with experience in cognition (Level B).
nitive status (Level B).
Section B: Recommendations for management
Recommendation A6
of MCI
Rationale
Recommendation B1
Although MCI is a high-risk state for progression to dementia,
Rationale
some patients with MCI remain stable and some improve.
Some cases of MCI are associated with reversible causes of
Some patients with MCI improve or remain stable rather than
cognitive impairment, including medication side eﬀects, sleep
progress. In addition, some cases of MCI are associated with
apnea, depression, and other medical conditions.e33 Patients
reversible causes of cognitive impairment, including medication
with MCI should undergo a medical evaluation for MCI risk
side eﬀects, general medical conditions, sleep disturbance, and
factors that may be treatable.
depression.e33 Because these risk factors are treatable and have
implications of their own, weaning patients from use of cog-
Recommendation
nitively impairing medications where feasible and treating risk
For patients diagnosed with MCI, clinicians should perform
factors that may contribute to cognitive impairment should be
a medical evaluation for MCI risk factors that are potentially
the ﬁrst steps in managing MCI, particularly because symp-
modiﬁable (Level B).
tomatic treatment options are limited for impaired cognition.
Recommendation A7
Recommendation
Rationale
For patients diagnosed with MCI, clinicians should wean
Because patients with MCI can improve, remain stable,
patients from medications that can contribute to cognitive
or progress cognitively, identifying biomarkers that can stratify
impairment (where feasible and medically appropriate) and
risk is expected to be particularly important for prognosis. The
treat modiﬁable risk factors that may be contributing (Level B).
6
Neurology | Volume 90, Number 3 | January 16, 2018
Neurology.org/N
Copyright ª 2017 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Recommendation B2
in MCI. Exercise also has general health beneﬁts and generally
limited risk.
Rationale
There are no FDA-approved medications for the treatment of
MCI. Moreover, there are no high-quality, long-term studies
Recommendation
identifying pharmacologic or dietary agents that either improve
For patients diagnosed with MCI, clinicians should recom-
cognition or delay progression in patients with MCI.
mend regular exercise (twice/week) as part of an overall ap-
proach to management (Level B).
Recommendation
Recommendation B6
For patients diagnosed with MCI, clinicians should counsel
Rationale
the patients and families that there are no pharmacologic or
dietary agents currently shown to have symptomatic cognitive
Because the concept of MCI may be poorly understood or
beneﬁt in MCI and that no medications are FDA-approved
distressing to patients and families, it is important to edu-
for this purpose (Level B).
cate patients and families regarding the diagnosis of MCI
and how it may progress to dementia but also how indi-
Recommendation B3
viduals with MCI can remain stable or improve. Because
Rationale
MCI may progress to dementia, and particularly because of
the lack of eﬀective pharmacologic therapy or any proven
Studies of cholinesterase inhibitors showed no beneﬁt on
methods to reduce the risk of progression of MCI to de-
cognitive outcomes or reduction in progression from MCI to
mentia, it is particularly important to educate patients
dementia, although some studies could not exclude an im-
with MCI regarding their diagnosis and prognosis at the
portant eﬀect. In addition to lacking eﬃcacy, side eﬀects of
MCI stage while they can still understand the discussion
cholinesterase inhibitors are common, including gastrointes-
and participate in planning, even though they may or may
tinal symptoms and cardiac concerns.e43
not progress. Because of the possibility of progression
to a dementia state where patients may no longer be able
Recommendation B3a
to participate in decision making, patients with MCI should
be encouraged to participate in long-term planning, in-
For patients diagnosed with MCI, clinicians may choose not
cluding topics such as advance directives, living wills, power
to oﬀer cholinesterase inhibitors (Level B).
of attorney designations, and ﬁnances, which are important
irrespective of progression.
Recommendation B3b
If clinicians choose to oﬀer cholinesterase inhibitors, they must
Recommendation
ﬁrst discuss with patients the fact that this is an oﬀ-label pre-
scription not currently backed by empirical evidence (Level A).
For patients diagnosed with MCI, clinicians should discuss
diagnosis and uncertainties regarding prognosis. Clinicians
Recommendation B4
should counsel patients and families to discuss long-term
planning topics such as advance directives, driving safety,
Rationale
ﬁnances, and estate planning (Level B).
Clinical trials provide an opportunity for interested patients to
participate in identifying or testing new treatment options,
Recommendation B7
which is of particular importance when no pharmacologic
Rationale
options are available.
Although there are no treatments for cognitive symptoms
in MCI, clinicians need to evaluate for and treat other
Recommendation
symptoms that can contribute to quality of life in MCI.
For patients diagnosed with MCI who are interested in
Behavioral/psychiatric symptoms are common in MCIe44–
e46
pharmacologic treatment, clinicians may inform these
and may be associated with greater functional impair-
patients of centers or organizations that can connect patients
mente47 and an increased risk of progression from MCI to
to clinical trials (e.g., subspecialty centers, Trial Match,
dementia.e48,e49
ClinicalTrials.gov) (Level C).
Recommendation B5
Recommendation
Clinicians should assess for behavioral and neuropsychi-
Rationale
atric symptoms in MCI and treat with both pharmaco-
Although long-term studies are unavailable, 6-month studies
logic and nonpharmacologic approaches when indicated
suggest a possible beneﬁt of twice-weekly exercise for cognition
(Level B).
Neurology.org/N
Neurology | Volume 90, Number 3 | January 16, 2018
7
Copyright ª 2017 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Recommendation B8
a particular use or purpose. The AAN assumes no re-
sponsibility for any injury or damage to persons or property
Rationale
arising out of or related to any use of this information or for
In patients with MCI, cognitive interventions may be bene-
any errors or omissions.
ﬁcial in improving measures of cognitive function. It is good
practice to oﬀer nonmedication approaches to care.
Conflict of interest
The American Academy of Neurology (AAN) is committed
Recommendation
to producing independent, critical, and truthful clinical prac-
In patients with MCI, clinicians may recommend cognitive
tice guidelines (CPGs). Signiﬁcant eﬀorts are made to mini-
interventions (Level C).
mize the potential for conﬂicts of interest to inﬂuence the
recommendations of this CPG. To the extent possible, the
AAN keeps separate those who have a ﬁnancial stake in
Suggestions for future research
the success or failure of the products appraised in the CPGs
and the developers of the guidelines. Conﬂict of interest forms
The guideline panel recommends (1) the use of consistent
were obtained from all authors and reviewed by an oversight
diagnostic criteria for MCI and dementia in clinical trials, to
committee prior to project initiation. The AAN limits the
improve the ability to apply and combine results; (2) the
participation of authors with substantial conﬂicts of interest.
inclusion of patient cohorts with speciﬁc biomarker data in
The AAN forbids commercial participation in, or funding of,
treatment studies targeted at speciﬁc pathologies (e.g., MCI
guideline projects. Drafts of the guideline have been reviewed
due to AD); (3) the use of outcome measures that are direct
by at least 3 AAN committees, a network of neurologists,
measures
of
clinically
meaningful
patient
outcomes
Neurology peer reviewers, and representatives from related
(i.e., development of dementia, reduction of ability to un-
ﬁelds. The AAN Guideline Author Conﬂict of Interest Policy
dertake activities of daily living or IADL, patient or caregiver
can be viewed at aan.com. For complete information on this
[or both] quality of life measures) or surrogate markers that
process, access the 2004 AAN process manual.7
have previously been shown to have a strong correlation with
such measures; (4) standardized reporting of trial design in
publications using CONSORT criteriae50; (5) study of MCI
Author contributions
thought to be secondary to AD and MCI related to other
Dr. Petersen: study concept and design, acquisition of data,
pathologies (e.g., vascular MCI, MCI related to Lewy body
analysis or interpretation of data, drafting/revising the man-
pathology); and (6) further study of early lifestyle and
uscript, critical revision of the manuscript for important in-
comorbidity modiﬁcations and the eﬀects of such changes on
tellectual content, study supervision. Dr. Lopez: study
the progression of MCI to diﬀerent dementia subtypes.
concept and design, analysis or interpretation of data,
drafting/revising the manuscript, critical revision of the
manuscript for important intellectual content. Dr. Armstrong:
Disclaimer
analysis or interpretation of data, drafting/revising the man-
Clinical practice guidelines, practice advisories, systematic
uscript, critical revision of the manuscript for important in-
reviews, and other guidance published by the American
tellectual content, study supervision. T.S.D. Getchius: study
Academy of Neurology (AAN) and its aﬃliates are assess-
concept and design, study supervision. Dr. Ganguli: study
ments of current scientiﬁc and clinical information provided
concept and design, analysis or interpretation of data,
as an educational service. The information (1) should not be
drafting/revising the manuscript, critical revision of the
considered inclusive of all proper treatments, methods of care,
manuscript for important intellectual content. Dr. Gloss:
or as a statement of the standard of care; (2) is not continually
analysis or interpretation of data, study supervision.
updated and may not reﬂect the most recent evidence (new
Dr. Gronseth: analysis or interpretation of data. Dr. Marson:
evidence may emerge between the time information is de-
study concept and design, analysis or interpretation of data,
veloped and when it is published or read); (3) addresses only
drafting/revising the manuscript, critical revision of the
the question(s) speciﬁcally identiﬁed; (4) does not mandate
manuscript for important intellectual content. Dr. Prings-
any particular course of medical care; and (5) is not intended
heim: analysis and interpretation of data, study supervision.
to substitute for the independent professional judgment of the
Dr. Day: analysis and interpretation of data, study supervision.
treating provider, as the information does not account for
Dr. Sager: study concept and design, analysis or interpretation
individual variation among patients. In all cases, the selected
of data, critical revision of the manuscript for important in-
course of action should be considered by the treating provider
tellectual content. Dr. Stevens: study concept and design,
in the context of treating the individual patient. Use of the
analysis or interpretation of data, drafting/revising the man-
information is voluntary. The AAN provides this information
uscript, critical revision of the manuscript for important in-
on an “as is” basis and makes no warranty, expressed or im-
tellectual content. Dr. Rae-Grant: analysis or interpretation of
plied, regarding the information. The AAN speciﬁcally dis-
data, drafting/revising the manuscript, critical revision of the
claims any warranties of merchantability or ﬁtness for
manuscript for important intellectual content.
8
Neurology | Volume 90, Number 3 | January 16, 2018
Neurology.org/N
Copyright ª 2017 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Acknowledgment
2.
Petersen RC. Mild cognitive impairment as a diagnostic entity. J Intern Med 2004;
256:183–194.
The authors thank Patricia J. Erwin of the Mayo Medical
3.
Petersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG, Kokmen E. Mild cognitive
Library for performing the literature searches; Julie Cox,
impairment: clinical characterization and outcome. Arch Neurol 1999;56:303–308.
4.
Huey ED, Manly JJ, Tang MX, et al. Course and etiology of dysexecutive MCI in
MFA, and Jonathan Dittman, MA, MFA, of the AAN for their
a community sample. Alzheimers Dement 2013;9:632–639.
work in copyediting the manuscripts and related documents;
5.
Petersen RC, Stevens JC, Ganguli M, Tangalos EG, Cummings JL, DeKosky ST.
and Shannon Merillat, MLIS, of the AAN for her work in
Practice parameter: early detection of dementia: mild cognitive impairment (an
evidence-based review): report of the Quality Standards Subcommittee of the
coordinating the project.
American Academy of Neurology. Neurology 2001;56:1133–1142.
6.
Lyketsos CG, Lopez OL, Jones B, Breitner J, DeKosky ST. A population-based study
of the prevalence of neuropsychiatric disturbances in dementia and mild cognitive
Study funding
impairment: results from the Cardiovascular Health Study. JAMA 2002;288:
1425–1483.
This practice guideline was developed with ﬁnancial support
7.
American Academy of Neurology. Clinical Practice Guideline Process Manual, 2004
from the American Academy of Neurology. Authors who
ed. [online]. St. Paul: The American Academy of Neurology. Available at: aan.com/
uploadedFiles/Website_Library_Assets/Documents/2.Clinical_Guidelines/4.
serve as AAN subcommittee members, methodologists, or
About_Guidelines/1.How_Guidelines_Are_Developed/2004%20AAN%20Process
employees, past or present (M.J.A., T.S.D.G., D.G., G.S.G.,
%20Manual.pdf. Accessed March 7, 2008.
8.
American Academy of Neurology. Clinical Practice Guideline Process Manual, 2011
T.P., G.S.D, A.R.G.), were reimbursed by the AAN for
ed. Available at: aan.com/Guidelines/Home/Development. Accessed April 12, 2012.
expenses related to travel to subcommittee meetings where
9.
Ganguli M, Dodge HH, Shen C, DeKosky ST. Mild cognitive impairment, amnestic
type: an epidemiologic study. Neurology 2004;63:115–121.
drafts of manuscripts were reviewed.
10.
Fei M, Qu YC, Wang T, Yin J, Bai JX, Ding QH. Prevalence and distribution of
cognitive impairment no dementia (CIND) among the aged population and the
Disclosure
analysis of socio-demographic characteristics: the community-based cross-sectional
study. Alzheimer Dis Assoc Disord 2009;23:130–138.
R. Petersen has served as a consultant for Roche Inc., Merck
11.
Bennett DA, Schneider JA, Bienias JL, Evans DA, Wilson RS. Mild cognitive im-
Inc., Genentech Inc., and Biogen Inc.; receives publishing
pairment is related to Alzheimer disease pathology and cerebral infarctions. Neurol-
ogy 2005;64:834–841.
royalties from Oxford University Press; performs clinical
12.
Anstey KJ, Cherbuin N, Christensen H, et al. Follow-up of mild cognitive impairment
procedures relating to mild cognitive impairment in his clin-
and related disorders over four years in adults in their sixties: the PATH through Life
Study. Dement Geriatr Cogn Disord 2008;26:226–233.
ical neurology practice; and receives research support from
13.
Di Carlo A, Lamassa M, Baldereschi M, et al. CIND and MCI in the Italian elderly:
the National Institute on Aging of the NIH. O. Lopez has
frequency, vascular risk factors, progression to dementia. Neurology 2007;68:
1909–1916.
been a consultant for Grifols Inc., Lundbeck, and Raman
14.
Hanninen T, Hallikainen M, Tuomainen S, Vanhanen M, Soininen H. Prevalence of
Technologies and has received grant support from the NIH.
mild cognitive impairment: a population-based study in elderly subjects. Acta Neurol
Scand 2002;106:148–154.
M. Armstrong serves on the Level of Evidence editorial board
15.
Louis ED, Schupf N, Manly J, Marder K, Tang MX, Mayeux R. Association between
for Neurology® (but is not compensated ﬁnancially) and
mild parkinsonian signs and mild cognitive impairment in a community. Neurology
2005;64:1157–1161.
serves as an evidence-based medicine methodologist for the
16.
Purser JL, Fillenbaum GG, Wallace RB. Memory complaint is not necessary for
American Academy of Neurology (AAN). T. Getchius was an
diagnosis of mild cognitive impairment and does not predict 10-year trajectories of
functional disability, word recall, or short portable mental status questionnaire limi-
employee of the AAN and has nothing to disclose. M. Ganguli
tations. J Am Geriatr Soc 2006;54:335–338.
has served on the data safety and monitoring board for
17.
Schonknecht P, Pantel J, Kruse A, Schroder J. Prevalence and natural course of aging-
Indiana University and on the advisory committee for Biogen
associated cognitive decline in a population-based sample of young-old subjects. Am J
Psychiatry 2005;162:2071–2077.
Inc. and has received research support from the National
18.
Artero S, Ancelin ML, Portet F, et al. Risk proﬁles for mild cognitive impairment and
Institute on Aging of the NIH. D. Gloss serves as an evidence-
progression to dementia are gender speciﬁc. J Neurol Neurosurg Psychiatry 2008;79:
979–984.
based medicine methodologist for the AAN. G. Gronseth
19.
Boyle PA, Wilson RS, Aggarwal NT, Tang Y, Bennett DA. Mild cognitive impairment:
serves as associate editor for Neurology, serves on the editorial
risk of Alzheimer disease and rate of cognitive decline [see comment]. Neurology
2006;67:441–445.
advisory board for Neurology Now, and is compensated by the
20.
Busse A, Hensel A, Guhne U, Angermeyer MC, Riedel-Heller SG. Mild cognitive
AAN for methodologic activities. D. Marson serves as a con-
impairment: long-term course of four clinical subtypes. Neurology 2006;67:
2176–2185.
sultant for and received royalties from Janssen Pharmaceut-
21.
Das SK, Bose P, Biswas A, et al. An epidemiologic study of mild cognitive impairment
icals and receives research support from the National Institute
in Kolkata, India. Neurology 2007;68:2019–2026.
22.
Lobo A, Lopez-Anton R, de-la-Camara C, et al. Non-cognitive psychopathological
on Aging for the NIH. T. Pringsheim has received ﬁnancial
symptoms associated with incident mild cognitive impairment and dementia, Alz-
reimbursement for travel to attend the Movement Disorder
heimer’s type. Neurotox Res 2008;14:263–272.
23.
Lopez OL, Kuller LH, Becker JT, et al. Incidence of dementia in mild cognitive
Society Meeting from Allergan Canada and Teva Canada
impairment in the Cardiovascular Health Study Cognition Study. Arch Neurol 2007;
Innovation and has received research support from Shire
64:416–420.
Canada Inc. and the Canadian Institutes of Health Research.
24.
Petersen RC, Roberts RO, Knopman DS, et al. Prevalence of mild cognitive impairment
is higher in men: The Mayo Clinic Study of Aging. Neurology 2010;75:889–897.
G. Day received honoraria for serving as faculty at the 2016
25.
Wilson RS, Schneider JA, Arnold SE, Tang Y, Boyle PA, Bennett DA. Olfactory
AAN annual meeting and holds stock (>$10,000) in ANI
identiﬁcation and incidence of mild cognitive impairment in older age. Arch Gen
Psychiatry 2007;64:802–808.
Pharmaceuticals (generic manufacturer). M. Sager, J. Stevens,
26.
Ganguli M, Chang CC, Snitz BE, Saxton JA, Vanderbilt J, Lee CW. Prevalence
and A. Rae-Grant report no disclosures relevant to the man-
of mild cognitive impairment by multiple classiﬁcations: the Monongahela-
Youghiogheny Healthy Aging Team (MYHAT) project. Am J Geriatr Psychiatry
uscript. Go to Neurology.org/N for full disclosures.
2010;18:674–683.
27.
Lopez OL, Jagust WJ, DeKosky ST, et al. Prevalence and classiﬁcation of mild cog-
Received February 24, 2017. Accepted in ﬁnal form September 22, 2017.
nitive impairment in the Cardiovascular Health Study Cognition Study: part 1. Arch
Neurol 2003;60:1385–1389.
28.
Shi Z, Zhang Y, Yue W, et al. Prevalence and clinical predictors of cognitive impair-
References
ment in individuals aged 80 years and older in rural China. Dement Geriatr Cogn
1.
Winblad B, Palmer K, Kivipelto M, et al. Mild cognitive impairment: beyond con-
Disord 2013;36:171–178.
troversies, towards a consensus: report of the International Working Group on Mild
29.
Guaita A, Vaccaro R, Davin A, et al. Inﬂuence of socio-demographic features and
Cognitive Impairment. J Intern Med 2004;256:240–246.
apolipoprotein E epsilon 4 expression on the prevalence of dementia and cognitive
Neurology.org/N
Neurology | Volume 90, Number 3 | January 16, 2018
9
Copyright ª 2017 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

impairment in a population of 70-74-year olds: the InveCe.Ab study. Arch Gerontol
35.
Hilal S, Ikram MK, Saini M, et al. Prevalence of cognitive impairment in Chinese:
Geriatr 2015;60:334–343.
Epidemiology of Dementia in Singapore study. J Neurol Neurosurg Psychiatry 2013;
30.
Anttila T, Helkala E, Viitanen M, et al. Alcohol drinking in middle age and subsequent
84:686–692.
risk of mild cognitive impairment and dementia in old age: a prospective population
36.
Kivipelto M, Helkala E, Hanninen T, et al. Midlife vascular risk factors and late-life mild
based study. BMJ 2004;329:539–542.
cognitive impairment: a population-based study. Neurology 2001;56:1683–1689.
31.
Barcelos-Ferreira R, Bottino C. Prevalence of amnestic mild cognitive impairment in
37.
Lee SB, Kim KW, Youn JC, et al. Prevalence of mild cognitive impairment and its
depressed and nondepressed elderly Brazilian community residents. Alzheimers
subtypes are inﬂuenced by the application of diagnostic criteria: results from the
Dement 2014;10(suppl):P907.
Korean Longitudinal Study on Health and Aging (KLoSHA). Dement Geriatr Cogn
32.
De Ronchi D, Palmer K, Pioggiosi P, et al. The combined eﬀect of age, education, and
Disord 2009;28:23–29.
stroke on dementia and cognitive impairment no dementia in the elderly. Dement
38.
Li X, Ma C, Zhang J, et al; on behalf of the Beijing Ageing Grain Rejuvenation
Geriatr Cogn Disord 2007;24:266–273.
Initiative. Prevalence of and potential risk factors for mild cognitive impairment
33.
Ding D, Zhao Q, Guo Q, et al. Prevalence of mild cognitive impairment in an urban
in community-dwelling residents of Beijing. J Am Geriatr Soc 2013;61:2111–2119.
community in China: a cross-sectional analysis of the Shanghai Aging Study. Alz-
39.
Miyamoto M, Kodama C, Kinoshita T, et al. Dementia and mild cognitive im-
heimers Dement 2015;11:300–309.e302.
pairment among non-responders to a community survey. J Clin Neurosci 2009;16:
34.
Gavrila D, Antunez C, Tormo MJ, et al. Prevalence of dementia and cognitive im-
270–276.
pairment in Southeastern Spain: the Ariadna study. Acta Neurol Scand 2009;120:
40.
Olazaran J, Valenti M, Frades B, et al. The Vallecas Project: a cohort to identify early
300–307.
markers and mechanisms of Alzheimer’s disease. Front Aging Neurosci 2015;7:181.
What’s Your Story? Submit a Video to 2018 Neuro Film Festival
Do you have a powerful story about the impacts of brain disease on yourself or a loved one? Are you studying neuroscience in
school or considering neurology as a career? Do you want to help raise awareness of and funding for critical brain disease
research? Or are you simply fascinated by the wonders of the brain? Share your unique story for the 2018 Neuro Film Festival.
Video submission deadline is March 2, 2018. Learn more at NeuroFilmFestival.com.
Neurology in the Spotlight at 2018 Annual Meeting in Los Angeles
Registration is now open for the totally flexible, dynamic 2018 Annual Meeting. We’ll be shining the spotlight on neurology
and what you need to excel in your career. Look for the latest science, education, and networking you won’t find anywhere else
when the biggest names in neurology and neuroscience convene in Los Angeles April 21 through 27. Learn more and register
now at AAN.com/view/AM18.
Neurology.org/N Offers Important Information to Patients and Their Families
The Neurology® Patient Page provides:
• A critical review of ground-breaking discoveries in neurologic research that are written especially for patients and their families
• Up-to-date patient information about many neurologic diseases
• Links to additional information resources for neurologic patients
All Neurology Patient Page articles can be easily downloaded and printed, and may be reproduced to distribute for educational purposes.
Click on the ‘Patients’ link on the home page (Neurology.org/N) for a complete index of Patient Pages.
10
Neurology | Volume 90, Number 3 | January 16, 2018
Neurology.org/N
Copyright ª 2017 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.